In the initial grant, our aim was to select candidate tumor suppressor genes by their loss of expression in mammary carcinomas compared with well matched normal mammary epithelial cells using differential display. This objective has now been achieved. We have identified and cloned more than 100 down-regulated genes in our breast cancer system. In addition we have investigated several candidate tumor suppressor genes of which maspin, which is a protease inhibitior, shows promise in both diagnostic and therapeutic applications. The first specific aim of this grant renewal is to test the hypothesis that maspin, a serpin (serine protease inhibitor) acts as a tumor suppressor through its interaction with the serine protease, tissue plasminogen activator. The structure of maspin, deduced from its sequence, does not support a strong prediction concerning the protease inhibitory activity of the protein, and until now its molecular mode of action has remained underfined. At the cellular level, however, we have shown that maspin inhibits invasion and motility in cell culture assays, and inhibits growth and metastasis in the nude mouse assay. By time-lapse video microscopy, we showed that motility is blocked for 12 hours when tumor cells are treated with maspin. Our purpose now is to define how tissue plasminogen activator contributes to these biological effects. We cloned and sequenced the promoter region of maspin, and by CAT analysis established the transcriptional regulation of maspin expression in both mammary andprostate cells. We propose now to identify the transcription factors responsible for differential expression in normal vs. Tumor cells of both tissues. We cloned and sequenced the mouse maspin and showed that it has 87 percent homology with human maspin and similar activity in inhibiting invasion and motility. It is proposed now to look for tumor suppressor activity in transgenic mice crossed with mice that express high frequencies of spontaneous mammary tumors; and to produce maspin knockout mice to study effects of maspin in development. The second aim is to utilize a grid system based on reverse Northern bloct to compare patterns of gene expression in the 100 down-regulated genes we have isolated by DD. Gene probes arrayed on grids will be hybridized with 32P labeled reverse transcribed single strand cDNAs from carcinoma cell lines and from patient speciments. The purpose is to identify patterns of coordinate expression characteristic of breast cancer and thereby to select genes of special interest for diagnostic and therapeutic application.